Change Details

Rewrite the final loop in vm_phys_enqueue_contig as a new function, vm_phys_enq_beg. The number of amd64 bytes in the loop shrinks from 100 to 70. Old: ``` 1d80: 48 c1 f9 03 sar $0x3,%rcx 1d84: 48 0f af cf imul %rdi,%rcx 1d88: 48 0f bd c9 bsr %rcx,%rcx 1d8c: 83 f1 c0 xor $0xffffffffffffffc0,%ecx 1d8f: 83 c1 41 add $0x41,%ecx 1d92: 83 c1 ff add $0xffffffffffffffff,%ecx 1d95: 88 4e 5c mov %cl,0x5c(%rsi) 1d98: 48 8d 46 10 lea 0x10(%rsi),%rax 1d9c: 48 c7 46 10 00 00 00 movq $0x0,0x10(%rsi) 1da3: 00 1da4: 48 63 d1 movslq %ecx,%rdx 1da7: 48 8d 14 52 lea (%rdx,%rdx,2),%rdx 1dab: 49 8b 5c d2 08 mov 0x8(%r10,%rdx,8),%rbx 1db0: 48 89 5e 18 mov %rbx,0x18(%rsi) 1db4: 49 8b 5c d2 08 mov 0x8(%r10,%rdx,8),%rbx 1db9: 48 89 33 mov %rsi,(%rbx) 1dbc: 49 89 44 d2 08 mov %rax,0x8(%r10,%rdx,8) 1dc1: 41 83 44 d2 10 01 addl $0x1,0x10(%r10,%rdx,8) 1dc7: b8 01 00 00 00 mov $0x1,%eax 1dcc: d3 e0 shl %cl,%eax 1dce: 48 98 cltq 1dd0: 48 6b c0 68 imul $0x68,%rax,%rax 1dd4: 48 01 c6 add %rax,%rsi 1dd7: 4c 39 ce cmp %r9,%rsi 1dda: 73 0c jae 1de8 <vm_phys_enqueue_contig+0x198> 1ddc: 4c 89 c9 mov %r9,%rcx 1ddf: 48 29 f1 sub %rsi,%rcx 1de2: 75 9c jne 1d80 <vm_phys_enqueue_contig+0x130> ``` New: ``` 1d80: 0f bd ca bsr %edx,%ecx 1d83: 88 4e 5c mov %cl,0x5c(%rsi) 1d86: 48 8d 7e 10 lea 0x10(%rsi),%rdi 1d8a: 48 c7 46 10 00 00 00 movq $0x0,0x10(%rsi) 1d91: 00 1d92: 48 8d 1c 49 lea (%rcx,%rcx,2),%rbx 1d96: 49 8b 44 de 08 mov 0x8(%r14,%rbx,8),%rax 1d9b: 48 89 46 18 mov %rax,0x18(%rsi) 1d9f: 49 8b 44 de 08 mov 0x8(%r14,%rbx,8),%rax 1da4: 48 89 30 mov %rsi,(%rax) 1da7: 49 89 7c de 08 mov %rdi,0x8(%r14,%rbx,8) 1dac: 41 83 44 de 10 01 addl $0x1,0x10(%r14,%rbx,8) 1db2: b8 01 00 00 00 mov $0x1,%eax 1db7: d3 e0 shl %cl,%eax 1db9: 48 98 cltq 1dbb: 48 6b c8 68 imul $0x68,%rax,%rcx 1dbf: 48 01 ce add %rcx,%rsi 1dc2: 29 c2 sub %eax,%edx 1dc4: 75 ba jne 1d80 <vm_phys_enqueue_contig+0x130> ```

Rewrite the final loop in vm_phys_enqueue_contig as a new function, vm_phys_enq_beg. The number of amd64 bytes in the loop shrinks from 91 to 70. Old: ``` 1e20: 48 89 d1 movq %rdx, %rcx 1e23: 48 29 f1 subq %rsi, %rcx 1e26: 48 c1 f9 03 sarq $0x3, %rcx 1e2a: 48 0f af cf imulq %rdi, %rcx 1e2e: 48 0f bd c9 bsrq %rcx, %rcx 1e32: 88 4e 5c movb %cl, 0x5c(%rsi) 1e35: 4c 8d 46 10 leaq 0x10(%rsi), %r8 1e39: 48 c7 46 10 00 00 00 00 movq $0x0, 0x10(%rsi) 1e41: 41 89 c9 movl %ecx, %r9d 1e44: 4f 8d 0c 49 leaq (%r9,%r9,2), %r9 1e48: 4e 8b 54 c8 08 movq 0x8(%rax,%r9,8), %r10 1e4d: 4c 89 56 18 movq %r10, 0x18(%rsi) 1e51: 4e 8b 54 c8 08 movq 0x8(%rax,%r9,8), %r10 1e56: 49 89 32 movq %rsi, (%r10) 1e59: 4e 89 44 c8 08 movq %r8, 0x8(%rax,%r9,8) 1e5e: 42 ff 44 c8 10 incl 0x10(%rax,%r9,8) 1e63: 41 b8 01 00 00 00 movl $0x1, %r8d 1e69: 41 d3 e0 shll %cl, %r8d 1e6c: 49 63 c8 movslq %r8d, %rcx 1e6f: 48 6b c9 68 imulq $0x68, %rcx, %rcx 1e73: 48 01 ce addq %rcx, %rsi 1e76: 48 39 d6 cmpq %rdx, %rsi 1e79: 72 a5 jb 0x1e20 <vm_phys_enqueue_contig+0x130> ``` New: ``` 1e20: 0f bd ca bsrl %edx, %ecx 1e23: 88 4e 5c movb %cl, 0x5c(%rsi) 1e26: 48 8d 7e 10 leaq 0x10(%rsi), %rdi 1e2a: 48 c7 46 10 00 00 00 00 movq $0x0, 0x10(%rsi) 1e32: 4c 8d 04 49 leaq (%rcx,%rcx,2), %r8 1e36: 4e 8b 4c c0 08 movq 0x8(%rax,%r8,8), %r9 1e3b: 4c 89 4e 18 movq %r9, 0x18(%rsi) 1e3f: 4e 8b 4c c0 08 movq 0x8(%rax,%r8,8), %r9 1e44: 49 89 31 movq %rsi, (%r9) 1e47: 4a 89 7c c0 08 movq %rdi, 0x8(%rax,%r8,8) 1e4c: 42 ff 44 c0 10 incl 0x10(%rax,%r8,8) 1e51: bf 01 00 00 00 movl $0x1, %edi 1e56: d3 e7 shll %cl, %edi 1e58: 48 63 cf movslq %edi, %rcx 1e5b: 48 6b f9 68 imulq $0x68, %rcx, %rdi 1e5f: 48 01 fe addq %rdi, %rsi 1e62: 29 ca subl %ecx, %edx 1e64: 75 ba jne 0x1e20 <vm_phys_enqueue_contig+0x130> ```

Rewrite the final loop in vm_phys_enqueue_contig as a new function, vm_phys_enq_beg. The number of amd64 bytes in the loop shrinks from 10091 to 70. Old: ``` 1d80: 48 c1 f9 031e20: 48 89 d1 sar $0x3, movq %rdx, %rcx 1d84: 48 0f af cf1e23: 48 29 f1 imul %rdi, subq %rsi, %rcx 1d88: 48 0f bd c91e26: 48 c1 f9 03 bsr %rcx, sarq $0x3, %rcx 1d8c: 83 f1 c01e2a: 48 0f af cf xor $0xffffffffffffffc0,%ecx imulq %rdi, %rcx 1d8f: 83 c1 411e2e: 48 0f bd c9 add $0x41,%ecx bsrq %rcx, %rcx 1d92: 83 c1 ff1e32: 88 4e 5c add $0xffffffffffffffff,%ecx movb %cl, 0x5c(%rsi) 1d95: 88 4e 5c1e35: 4c 8d 46 10 mov %cl,0x5c leaq 0x10(%rsi), %r8 1d98: 48 8d1e39: 48 c7 46 10 00 00 00 00 lea movq $0x0, 0x10(%rsi),%rax 1d9c: 48 c7 46 10 00 00 00 movq $0x0,0x10(%rsi)1e41: 41 89 c9 movl %ecx, %r9d 1da3: 00 1da4: 48 63 d11e44: 4f 8d 0c 49 movslq %ecx,%rdxleaq (%r9,%r9,2), %r9 1da7: 48 8d 14 52 1e48: 4e 8b 54 c8 08 lea (%rd movq 0x8(%rax,%rdx,2),%rdx9,8), %r10 1dab: 49 8b 5c d2 081e4d: 4c 89 56 18 mov 0x8( movq %r10,%rdx,8),%rbx 0x18(%rsi) 1db0: 48 89 5e 181e51: 4e 8b 54 c8 08 mov %rbx,0x18(%rsi)q 0x8(%rax,%r9,8), %r10 1db4: 1e56: 49 8b 5c d2 089 32 mov 0x8(%r10,%rdx,8),%rbx movq %rsi, (%r10) 1db9: 48 89 331e59: 4e 89 44 c8 08 mov %rsi,(%rbxq %r8, 0x8(%rax,%r9,8) 1dbc: 49 891e5e: 42 ff 44 d2 08c8 10 mov %rax,0x8(%r10,%rdx incl 0x10(%rax,%r9,8) 1dc1: 41 83 44 d2 10 011e63: 41 b8 01 00 00 00 addl $0x1,0x10(%r10,%rdx,8) movl $0x1, %r8d 1dc7: b8 01 00 00 001e69: 41 d3 e0 mov $0x1,%eax shll %cl, %r8d 1dcc: d3 e01e6c: 49 63 c8 shl %cl,%eaxmovslq %r8d, %rcx 1dce: 48 98 cltq 1e6f: 48 6b c9 68 1dd0: 48 6b c0 68 imul q $0x68,%rax,%rax %rcx, %rcx 1dd4: 1e73: 48 01 c6e add %rax, addq %rcx, %rsi 1dd7: 4c1e76: 48 39 ced6 cmp %r9, cmpq %rdx, %rsi 1dda: 73 0c jae 1de8 <vm_phys_enqueue_contig+0x198> 1ddc: 4c 89 c9 mov %r9,%rcx 1ddf: 48 29 f1 sub %rsi,%rcx 1de2: 75 9c jne 1d8e79: 72 a5 jb 0x1e20 <vm_phys_enqueue_contig+0x130> ``` New: ``` 1d80: 1e20: 0f bd ca bsr bsrl %edx, %ecx 1d83: 1e23: 88 4e 5c mov movb %cl, 0x5c(%rsi) 1d86: 1e26: 48 8d 7e 10 lea leaq 0x10(%rsi), %rdi 1d8a: 1e2a: 48 c7 46 10 00 00 00 00 movq $0x0, 0x10(%rsi) 1d91: 001e32: 4c 8d 04 49 leaq (%rcx,%rcx,2), %r8 1d92: 48 8d 1c 49 1e36: 4e 8b 4c c0 08 lea (%rc movq 0x8(%rax,%rcx,2),%rbx8,8), %r9 1d96: 49 8b 44 de 081e3b: 4c 89 4e 18 mov 0x8(%r14,%rbx,8),%rax movq %r9, 0x18(%rsi) 1d9b: 48 89 46 181e3f: 4e 8b 4c c0 08 mov q 0x8(%rax,0x18(%rsi)%r8,8), %r9 1d9f: 1e44: 49 8b 44 de 089 31 mov 0x8(%r14,%rbx,8),%rax movq %rsi, (%r9) 1da4: 48 89 301e47: 4a 89 7c c0 08 mov %rsi,q %rdi, 0x8(%rax,%r8,8) 1da7: 49 89 7c de 081e4c: 42 ff 44 c0 10 mov %rdi,0x8(%r14,%rbx incl 0x10(%rax,%r8,8) 1dac: 41 83 44 de 10 011e51: bf 01 00 00 00 addl $0x1,0x10(%r14,%rbx,8) movl $0x1, %edi 1db2: b8 01 00 00 001e56: d3 e7 mov $0x1,%eax shll %cl, %edi 1db7: d3 e01e58: 48 63 cf shl %cl,%eaxmovslq %edi, %rcx 1db9: 48 98 cltq 1e5b: 48 6b f9 68 1dbb: 48 6b c8 68 imul q $0x68,%rax,%rcx %rcx, %rdi 1dbf: 1e5f: 48 01 cfe add %rcx, addq %rdi, %rsi 1dc2: 1e62: 29 c2a sub %eax, subl %ecx, %edx 1dc4: 1e64: 75 ba jne 1d8 jne 0x1e20 <vm_phys_enqueue_contig+0x130> ```